Abstract: To accurately identify the stable, expansion, and loss portions of mangrove spatial extent, this study proposed a remote sensing identification method for monitoring mangrove spatial extent changes by combining the accurate identification capability of the random forest (RF) algorithm with high-precision advantages of vegetation indices to reflect the mangrove status. Using a collection of Landsat series images of Guangxi coastal zone in years of 2010 and 2020. Multiple spectral indices, including the normalized difference vegetation index (NDVI), modified normalized difference water index (MNDWI), normalized difference water index (NDWI), and mangrove vegetation index (MVI) for the years 2010 and 2020, along with topographic elevation, slope, and the MVI change difference index, were used as classification features, and RF algorithm was applied to identify mangroves. Results indicate that firstly, using spatiotemporal change indices as classification features and the RF algorithm for mangrove identification effectively overcomes the influence of periodic tidal inundation and spectral similarities between mangroves and other vegetation. The F1 scores for 2010 and 2020 identifications are 0.981 and 0.977, respectively. Besides, the sample analysis indicates that the fluctuations in MVI can effectively reflect the changes of mangrove spatial extent. To address potential biases resulting from independently conducted mangrove identifications in different time periods, the construction of the MVI difference index is employed to assist in mangrove identification, which effectively ensures the accuracy of spatial extent change analysis. Results demonstrate that this identification method significantly improves the accuracy of remote sensing monitoring of mangrove spatial extent changes, contributing to the conservation and management of mangrove ecosystems.